[clang] [CIR][CUDA] Add offload.printf op and CUDA target for it (PR #230517)

[Konstantinos Parasyris via cfe-commits]

https://github.com/koparasy approved this pull request.

LGTM, can we add a test for `-triple spir64 -x hip -fcuda-is-device`. I believe you widened the case I mention below:

https://github.com/llvm/llvm-project/pull/230517